The pitfalls of exaggeration: molecular and morphological evidence suggests Kaliana is a synonym of Mesabolivar (Araneae: Pholcidae) JONAS J. ASTRIN1, BERNHARD MISOF & BERNHARD A. HUBER2 Zoologisches Forschungsmuseum Alexander Koenig, Adenauerallee 160, D-53113 Bonn, Germany. Corresponding author. E-mail: [email protected]; [email protected] Abstract When the Venezuelan genus Kaliana Huber, 2000 was described, it was based on a single male specimen that was mor- phologically unique among pholcid spiders, especially in its extremely exaggerated male genitalia. The morphology of the recently discovered female suggests a close relationship with Mesabolivar González-Sponga, 1998. Using molecular sequences (mitochondrial CO1, 16S, and nuclear 28S) of Kaliana yuruani Huber, 2000 and 53 other pholcid taxa (152 sequences, 19 of them sequenced in this study) in a Bayesian and a maximum parsimony approach, we show that Kaliana is not sister group of, but nested within the species-rich South American genus Mesabolivar. Therefore, we argue that Kaliana is a junior synonym of Mesabolivar (Mesabolivar yuruani, n. comb.). Complementing previous stud- ies on pholcid phylogeny, we also present evidence for a close relationship between Mesabolivar and Carapoia, support the synonymy of Anomalaia and Metagonia with molecular data, support the monophyly of 'ninetines' and question the recently postulated position of Priscula as nested within the New World clade. Key words: pholcid spiders, subfamily-level groups, Metagonia, Carapoia, Priscula, beta-taxonomy, phylogeny Introduction There seems to be a tendency for taxonomists to create new genera for highly ‗aberrant‘ species. For example, when the first spider species with directionally asymmetric male genitalia was discovered, a new genus was erected for it (Anomalaia González-Sponga, 1998). Subsequent morphological studies strongly suggested that Anomalaia mariguitarensis is just an unusual representative of the widespread and species-rich Neotropical genus Metagonia Simon, 1893 (Huber 2000, 2004). Our focus here is on a similar case. The genus Kaliana Huber, 2000 was established for a single and extremely unusual male specimen from Venezuela. The procur- sus, a male genital structure of pholcid spiders, is in this species about six times as long as in other representa- tives of the family. Other autapomorphic characters include the shape of the eye turret, the modifications of the clypeus, and the armature of the chelicerae. During an expedition to Venezuela in 2004, further specimens of Kaliana yuruani Huber, 2000 were col- lected, and a simple but unique female character—the median pocket (Fig. 4; see also Huber 2006)—sug- gested an affinity to the widespread and species-rich South American genus Mesabolivar González-Sponga, 1998. This median pocket, located ventrally on the female genital plate (epigynum), has in fact been the only morphological synapomorphy of Mesabolivar (Huber 2000). Morphology thus suggested Kaliana to be either the sister group of Mesabolivar or to be nested within (and thereby a synonym of) Mesabolivar—without being able to falsify one of these two hypotheses. Here we use molecular evidence to test these phylogenetic hypotheses. We also use molecular data to highlight close evolutionary ties between Mesabolivar and Carapoia ssssss Accepted by C. Kropf: 8 Oct. 2007; published: 26 Nov. 2007 17 González-Sponga, to confirm the synonymy of Anomalaia and Metagonia, and to question the recently postu- lated position of Priscula Simon as nested within the New World clade. As for the subfamily-level clades pro- posed by Huber (2000), our data strongly corroborates monophyly of the New World clade (a large group of genera endemic to the New World); it further suggests inclusion of certain ‗holocnemines‘ in the ‗pholcines‘ (a suggestion that is strongly opposed by morphological evidence), and coincides with Bruvo-Mađarić et al. (2005) in rejecting ‗holocnemines‘ as a para- or polyphyletic group (both 'holocnemines' and 'pholcines' have a world-wide distribution; the former contain many unusually large forms while the species-rich 'pholcines' are best known for the synanthropic spider Pholcus phalangioides). This is the first molecular phylogenetic study to include more than one representative of the 'ninetines' (mostly tiny, ground-living pholcids) and thus the first one to test ninetine monophyly using genetic characters. A large part of the taxa and sequences used in this study has been analyzed before; our sampling mostly emerges as a synthesis of the studies of Astrin et al. (2006) and Bruvo-Mađarić et al. (2005). However, the first study (providing about two thirds of the sequences) was concerned with molecular alpha-taxonomy rather than with phylogeny, and the second differs in the analytical methods used (see below). Material and methods Genetic markers and taxon sampling Our principal focus lay on mitochondrial genes, cytochrome c oxidase subunit 1 (CO1) and the ribosomal large subunit (16S) (treated together for their provenance from a single linkage unit), thus complementing the data of Bruvo-Mađarić et al. (2005), who present results for these genes almost exclusively in combination with nuclear DNA (nDNA). The CO1 and 16S fragments used here are short due to the fact that most of them stem from a taxonomic study (Astrin et al. 2006). However, the combination of both markers produced some interesting and robust results. We used CO1 and 16S sequences from 60 specimens: 48 were taken from Astrin et al. (2006), nine from Bruvo-Mađarić et al. (2005) and three (cf. Table 1; see Table 2 for voucher, collecting data) were added for this study. The 42 represented pholcid species belong to 18 genera and all four currently recognized subfam- ily-level taxa (Huber 2000). Two non-pholcid outgroup species were used: a filistatid (Filistatidae being the putative sister family to all Haplogynae; Coddington and Levi 1991), and a representative of Diguetidae (together with Plectreuridae, the putative sister to Pholcidae). We included multiple specimens per species in cases in which the conspecific haplotypes were not identical. Some genera (especially the Neotropical endem- ics Mesabolivar González-Sponga, and Metagonia) are here represented by several species. Table 1 lists taxon names and GenBank accession numbers along with the country of origin. It is important to stress the fact that for the mitochondrial partition, we only chose specimens for which CO1 and 16S sequences were both avail- able since we rejected the option to enlarge the sampling at the cost of having to code entire partitions as miss- ing (see below). In order to consider focal taxa through another, independent and more conserved marker, we also included a number of nuclear 28S sequences. Therefore, we selected and sequenced 13 taxa (see Tables 1 and 2). We incorporated these into a 28S dataset with 22 pholcid and two outgroup taxa (plus one dubious sequence, see Discussion) that we obtained from Bruvo-Mađarić et al. (2005; cf. Table 1). Of these 38 sequences, 24 belong to taxa also present in the mtDNA dataset (i.e. 14 taxa were new, since one sequence per species was used for 28S). Finally, we concatenated the mitochondrial and the nuclear partitions—but again, only considering the (24) species in which both were available. For this combined approach, we had to use sequences obtained from different specimens in some species. 18 • Zootaxa 1646 © 2007 Magnolia Press ASTRIN ET AL. TABLE 1. List of analyzed specimens and GenBank (www.ncbi.nih.gov) accession numbers. Accession numbers start- ing with "AY": Bruvo-Mađarić et al. (2005); starting with "DQ": Astrin et al. (2006); accession numbers in boldface: specimens sequenced in this study (cf. Table 2). Informal higher level taxonomy: hol = holocnemines, nin = ninetines, NWC = New World clade, phol = pholcines. Taxon group CO1 16S 28S Artema atlanta Walckenaer, 1837 hol AY560771 AY560663 -- Artema atlanta Walckenaer, 1837 hol DQ667854 DQ667748 -- Carapoia paraguaensis González-Sponga, 1998 NWC DQ667855 DQ667749 DQ667839 Carapoia ubatuba Huber, 2005 NWC DQ667856 DQ667750 DQ667840 Ciboneya antraia Huber & Pérez, 2001 NWC AY560794 AY560665 AY560732 Coryssocnemis simla Huber, 2000 NWC DQ667858 DQ667753 -- Coryssocnemis simla Huber, 2000 NWC DQ667859 DQ667752 -- Crossopriza lyoni (Blackwall, 1867) hol AY560775 AY560667 DQ667841 Holocnemus pluchei (Scopoli, 1763) hol -- -- DQ667842 Ibotyporanga naideae Mello-Leitão, 1944 nin DQ667852 DQ667837 DQ667843 "Kaliana yuruani" = Mesabolivar yuruani (Huber, 2000) NWC DQ667860 DQ667754 DQ667844 Mecolaesthus longissimus Simon, 1893 NWC DQ667861 DQ667756 AY560736 Mesabolivar aurantiacus (Mello-Leitão, 1930) NWC AY560779 AY560670 AY560735 Mesabolivar aurantiacus (Mello-Leitão, 1930) NWC AY560778 AY560669 -- Mesabolivar aurantiacus (Mello-Leitão, 1930) NWC DQ667862 DQ667757 -- Mesabolivar brasiliensis (Moenkhaus, 1898) NWC -- -- AY560738 Mesabolivar cyaneotaeniatus (Keyserling, 1891) NWC AY560781 AY560671 AY560739 Mesabolivar cyaneotaeniatus (Keyserling, 1891) NWC DQ667866 DQ667760 -- Mesabolivar eberhardi Huber, 2000 NWC DQ667870 DQ667763 DQ667845 Mesabolivar eberhardi Huber, 2000 NWC DQ667872 DQ667764 -- Mesabolivar luteus (Keyserling, 1891) NWC DQ667873 DQ667766 DQ667846 Mesabolivar sp. 1 NWC DQ667874 DQ667767 -- Mesabolivar sp. 2 NWC DQ667875 DQ667769 -- Mesabolivar sp. 3 NWC DQ667877 DQ667771 -- Mesabolivar sp. 3 NWC DQ667876 DQ667770 -- Mesabolivar sp. 4 (M. sp. 6 in Astrin et al. 2006) NWC DQ667882 DQ667775